Professional Documents
Culture Documents
The Prepositional Inventory of Languages A Factor That Affects
The Prepositional Inventory of Languages A Factor That Affects
19-35, 1996
Pergamon Copyright ~ 1996 Published by Elsevier Science Lid
Printed in Great Britain. ,MI rights reserved
0388-0001/96 $15.00 + 0.00
S0388-0001(96)00005-8
ABSTRACT
Miller and Johnson-Laird's Language and Perception (1976) can be considered a starting point
of the psychological and linguistic investigation into deictic and intrinsic spatial reference, i.e.
into the use and the comprehension of spatial prepositions and their corresponding spatial
concepts. Important factors that affect the choice of either the deictic or the intrinsic frame of
reference have been suggested to be the discourse context, the static vs, dynamic
characteristics of the described situation, the individual cognitive style, and the intrinsic
orientation of the reference object, However, the explanatory determination of people's actual
use of spatial expressions is far from being satisfactory.
We conducted a series of experiments in German, Dutch, French, Italian and English in order
to show that four factors systematically interact when people interpret spatial expressions: (1)
the reference object being intrinsically oriented or not; (2) the preposition that is used; (3) the
social characteristics of the discourse situation; (4) the prepositional inventory of the language
at issue, as regards its spatial and temporal prepositions. Particularly, factors (3) and (4) allow
for a more comprehensive explanation of the empirical results, compared to the hitherto
proposed approaches. In all five languages, some communicative fuzziness, or ambiguity,
remains; however, it occurs in different cases, depending on the prepositional inventories of the
respective languages.
KEYWORDS
Spatial reference; spatial communication; prepositions; deixis; psycholinguistics.
INTRODUCTION
Space is a basic human experience (Gosztonyi, 1976). Therefore, people often communicate,
among other things, about their spatial environment (see Vater, 1991, for a general survey). In
our paper, communication about a quite simple case of spatial facts is dealt with: An object is
in a certain place, or it should be in a certain place. Objects, in a broad range of senses, include
all material things in the world, i.e. organisms as well as physical things. We restrict ourselves
19
20 .1. GRABOWSKI and P. WEISS
to material objects only for simplification; as Jackendoff (1983) points out, states and events
might also be assigned to places.
Basically, there are two reasons why the place of an object is referred to in communication
(Hemnann and Grabowski, 1994): (1) The speaker intends to emphasize an object in order to
direct the hearer's attention to it. This objective can be accomplished by different means: The
speaker can simply name the object (or the category of objects it belongs to) ("the ball"); he or
she can--in a more or less specified way--describe some of the attributes of an object ("the
red, round thing"); or the speaker refers to the place where the intended object is ("the ball on
the carpet"). Thus, if the speaker's goal is to focus on an object, the name of the object, its
description and its localization serve the same function. (2) The speaker intends to emphasize a
place where the object in question is presently located; or where it should be; or where it had
previously been; or where it had never been before; or where it will be in future; or where it
never should be; etc.
Regardless of the communicative purpose that is intended through the specification of the
place of an object: The common characteristic of localizing utterances is a spatial relation being
established between an intended object and a reference object, or between the place of an
intended object and the place occupied by a reference object. (It is an open question whether
objects or places take the arguments of spatial relations; cf. Habel, 1989.)
An appropriate means in order to specify spatial relations between objects is the use of
prepositions. Spatial prepositions can be subdivided according to the type of relation they
express. Prepositions like /n or near refer to the topological structure of space (inclusion,
contact, and proximity); beside, behind, above, and many other prepositions refer to the
dimensional structure of space (cf. Jackendoff and Landau, 1991; Klein, 1994; Landau and
Jackendoff, 1993; Retz-Schmidt, 1988). There are also prepositions that combine topological
and dimensional features, like on which is semantically defined by the conjunction of contact
(topological) and above (dimensional). In the following, we restrict ourselves to dimensional
prepositions (or projective prepositions, as called by Herskovits, 1986); therefore, we are
concerned with the field of secondary spatial deixis (Ehrich, 1985). In the studies reported
below, only spatial relations on the primary horizontal axis, i.e. the sagittal axis, are considered
which are expressed by the English prepositions in front of and behind, and by the German
prepositions vor and hinter, respectively. However, implicit treatment is also given to spatial
relations on the secondary horizontal axis, i.e. the transversal axis, which is expressed by on
the left and on the right, or links and rechts, respectively. For spatial relations on the vertical
line, however, some special reflections would be needed (Friederici, 1989; Leveit, 1984, 1986).
Firstly, we will explain why ambiguity arises with the interpretation of the prepositions that
refer to relations on the first horizontal axis, or with the interpretation of utterances that
contain these prepositions, respectively. Secondly, we will outline some attempts that have
already been introduced in order to explain the conditions in which the prepositions are used in
the one or in the other way. Thirdly, we will present some of our own experiments on the
prepositional use in German. The results show that there is, in fact, substantial ambiguity when
the prepositions are interpreted. There are at least three factors which determine the
interpretation of German vor [in front o.1] and hinter [behind]; however, they altogether do not
allow for the complete prediction of the hearers' interpretations. In the last section, we
introduce the hypothesis that part of the existing ambiguity is based on the fact that German
belongs to a particular type of languages, as regards its prepositional inventory. Further
experiments in different languages show that this hypothesis proves true. From this, it will be
concluded what extensions are needed to reach a comprehensive theory of the use of spatial
prepositions that is not restricted to a particular type oflangnages.
COMPREHENSION OF SPATIAl, PREPOSITIONS 21
The linguistic expressions for the dimensional subspaces are called the system of secondary
spatial deixis (Ehrich, 1985; here, we restrict ourselves to prepositional expressions,
neglecting adverbs and adjectives.) The traditional semantic description of these expressions
makes use of a localizing function; this function localizes the object that is to be localized in a
particular subspace of the reference object. Thus, the function uses the topological inclusion
and the subdivision of the region of the reference object into subspaces (Wunderlich and
Herweg, 1991). According to this approach, e.g., behind the church is semantically described
as IN the BEHIND-subspace of the church, or 1N the subspace that adjoins the back side of
the church, where the back side of the church is the side that points to the negative pole of the
primary horizontal axis, and where 1N is the topological inclusion. The meanings of the other
dimensional prepositions can be defined similarly. This way of semantically defining
prepositional meanings does not reflect our opinion; we prefer, for several reasons, to
reconstruct the meanings of spatial prepositions in terms of the perceptual accessibility of
objects and not in terms of the inclusion in subspaces adjacent to the reference object
(Grabowski, i. prep.). However, we can leave it at that for the following. It should be noticed
that, in the traditional semantical view, the comprehension of a prepositional expression and
the identification of the intended subspace presupposes the identification of the corresponding
side of the reference object.
When the human body is regarded as the primary origo that dimensions space, no problems
arise with the determination of the front side, the back side, etc., because of the
aforementioned asymmetries of the human body. However, there are many cases where the
Ego is no reliable candidate to serve as an anchor of spatial relations. Besides egocentricity, we
conceive ourselves to move through the word, and we conceive the world to be a world of
stable relations between objects. With our own movement, we continuously change our
position and the relative orientation of our egocentric body axes, particularly the horizontal
axes: You turn around, and what has been in front of you is now on your left. Moreover, we
know that every other individual constitutes an egocentric origo as well that defines the spatial
axes from his or her perspective. Therefore, it is useful to establish the spatial position of an
object not only with respect to one's own egocentric origo, but to establish a relation to
another object that is presumably more stable with respect to its spatial position and
orientation. For that, it is necessary to determine the sides of the reference object; here, its
front and back. However, this leads to three problems: (a) There are several principles
according to which the front, or the back, respectively, can be assigned to an object; (b) in
some cases, the alternative principles of front and back assignment compete, i.e. they lead to
conflicting assignments; (c) in most cases, it is not possible to deduce from spatial expressions
which principle has been used.
The two main principles that speakers and hearers use to assign the front and back of objects--
in order to make the establishment of a dimensional relation possible--are the extrinsic and the
intrinsic frame of reference (Carlson-Radvansky and Irwin, 1993; Ehrich, 1985; Grabowski,
1994; Levelt, 1986; Retz-Schmidt, 1988). Within the intrinsic frame of reference, the speaker
refers to the inherent spatial properties and axial asymmetries of the reference object. As
described above, these properties exist for human beings and for all other living things that can
move, but also for physical objects that are used in a characteristic orientation, e.g., cars and
vehicles in general, clothes, clocks, closets and desks. Within the extrinsic frame of reference,
fronts are assigned to objects by characteristics of the situation context. Here, different
principles exist (Wunderlich, 1982); we only refer to a particular principle which is often
confronted with the intrinsic system: the deictic frame of reference. In the deictic system, the
speaker's, or the observer's, position and orientation is used in order to determine the front of
an object. How does the deictic principle work? In our culture, it follows the principle of
COMPR~tIt~NSION 01, SPA I‘IAI, PRIIf’OSI’I‘IONS 23
canonical encounter (Clark, 1973), which is also called the facing strategy: The side of an
object that faces the observer, or the speaker, is attributed its front. Within other cultures (Hill,
1982), or within alternative systems of the extrinsic frame in our culture, e.g., when an object
is moving, the aligning strategy is used for deictic front assignment; here, the reference object
is conceived as being co-oriented with the observer, thus ‘looking’ into the same direction as
the observer and having its front at the side that does not face the observer. Figure 1 shows the
deictic use, based on the facing strategy as it is predominantly used with static objects, and the
intrinsic use of in front of and behind.
6 behind
RO
in front of
f
Fig. 1. The use of in front of and behind in the deictic (left box)
and in the intrinsic (right box) frame of reference. In the
deictic case the observer (Obs) induces the front of the
reference object (RO) by facing it; this allows for the
determination of the orientation of the reference object
on the primary horizontal axis.
It is possible to formally define the uses of in front of and behind in terms of an Euclidean
spatial structure (cf. Grabowski, 1994; Klein, 1994).
In the deictic system, the observer, or speaker, is the origin 0 of a coordinate system; his or
her line of vision constitutes the positive segment of the ordinate, with y(0) = 0. Let L be the
localized object and RO the reference object with y(L) > 0 and y(R0) > 0; i.e., the related
objects are in the origin’s line of vision. L is in front of RO, if y(L) < y(R). L is behind RO, if
Y(L) ’ Y(RO).
In the intrinsic system, which is independent of an observer, the reference object RO
constitutes the origin with y(R0) = 0. L is in front of RO, if y(L) > 0. L is behind RO, if
y(L) < 0. In this formal definition it is disregarded to what extent the use of inpont of and
behind is tolerant of lateral deviations, i.e. which relations of ly(L) - y(RO)I to Ix(L) - x(RO)I
can be managed with the use of infvont of and behind and which call for the use of Ieft and
right or for combined expressions of both dimensions.
The intrinsic system can only be used in cases where the reference object is intrinsically
oriented. However, the problem of ambiguity arises from the fact that in these cases the deictic
system can also be used and that the deictic and the intrinsic uses of infront of and behind lead
to conflicting interpretations, as shown in Fig. 1. Language offers means to unambiguously
indicate which system has been used: Expressions like from my point of view or from your
LX 18/1-2-B
24 .1. (II~,AI3()WSKIzlndP. WI,IIS.S
point of view can only be meant, and understood, deictically; phrases like on the left hand side
or at the front of can only be meant, and understood, intrinsically. However, in everyday
communication people almost without exception refrain from using these means (Grabowski
and Miller, i. prep.). Therefore, localizing utterances that refer to an intrinsically oriented
reference object are in most cases ambiguous. This ambiguity is illustrated in Fig. 2 where the
spatial positions of both cats can be appropriately expressed by both sentences The cat is in
front of the convertible, and The cat is on the left of the convertible.
Fig. 2. ]he cat is in front of the convertible.- The cat is" on the
left of the convertible.
Ehrich (1985) examines room descriptions and shows that the integration of separate groups of
objects into more extensive spatial arrangements is facilitated when a strategy, i.e. deictic or
intrinsic reference, is kept up through the entire discourse (el. Linde and Labov, 1975; von
Stutterheim and Carroll, 1993). However, spatial expressions are not always embedded in an
extensive spatial discourse context as it is the case with room descriptions or directions; they
sometimes appear separately, or in a non-spatial context, where the hearer can not refer to a
previously chosen strategy of reference in order to interpret a spatial expression either
deictically or intrinsically.
Wunderlich (1981) points out that the static vs. dynamic character of a situation in which an
object is localized--expressed by the use of static or dynamic verbs--influences the
interpretation of spatial relations between objects. Deictic interpretation is preferred for spatial
relations that include motion verbs (e.g., Put the lamp to the right of the sofa.), whereas
intrinsic interpretation is preferred for spatial expressions that include static verbs (e.g., The
lamp shouM remain on the right of the sofa.) (In German, no differences arise between the use
of prepositions in the context of dynamic or static verbs, as opposed to the change from to the
right to on the right in the English example.) However, there is no deterministic
correspondence between the chosen frame of reference and the static vs. dynamic factor;
Wunderlich merely found correspondences that range from 70 to 80 per cent.
Levelt (1982) points to individual styles and preferences in the choice of the deictic or intrinsic
frame of reference. However, we found in our experiments that inconsistent interpretations
arise between individuals as well as within individuals (Grabowski, i. prep.). Furthermore, the
assumption of individual styles does not appear to be plausible at all: Imagine that somebody
developed such a style of either constant deictic or constant intrinsic preference. It seems
reasonable that this localizing strategy is applied to the interpretation of utterances as well as
to the production of one's own utterances. However, we do not select our communication
partners according to their style of spatial reference; spatial communication must succeed
between any partners. If the 'deictics' and the 'intrinsics' were equally distributed in the
population, maximum fuzziness of communication would arise; i.e., only 50 per cent of the
utterances that can be interpreted either deictically or intrinsically would be interpreted by
hearers in the same way than they had been intended by speakers; that is chance level. On the
other hand, if one of both styles would be predominant in the population, one wonders why the
members of the smaller group have their strategy of spatial reference cultivated as a style in the
face of their frequent communicative failure. Participants in experiments are generally prone to
develop constant patterns of reaction when they are repeatedly exposed to similar tasks; these
patterns may look like stable habits of general behavior, although they merely reflect a task-
dependent response set.
Each of the mentioned approaches, as well as further approaches that we did not mention here,
consider one aspect that contributes to the prediction of the deictic vs. intrinsic interpretation
of spatial expressions. However, even if they are taken together they do not allow for a
comprehensive and satisfactory answer to the question of how dimensional prepositions are
used. In the experiments reported next, we will show that in German, in fact, considerable
ambiguity exists when the prepositions vor [in front o31 and hinter [behind] are interpreted,
and we will examine the conditions in which ambiguity arises. Then, we will introduce several
factors that have been experimentally considered in order to clarify the problem.
26 ,I. (ilIAt~OWSKI and P. Wt,21SS
I ~ oar
subspace3
yb:,.~.l.
I II~ c'r
: ror
Fig. 3. The setting that was used for the experiments on the use
of vor [in front of] and hinter [behind].
Firstly, we embedded the basic pattern of our setting in a private and informal situation. Here,
participants were instructed as follows (instructions are given in translation): "Imagine you are
giving a friend a lilt home. But you don't know exactly where he lives. He says: Could you
C()MPRI.~It]NSION OF SPATIAl, I~RI,It~()SITIONS 27
please drop me off in front of [vor]/behind [hinter] the yellow beetle/the tree? Where would
you stop your car?"
Considering the state of the art outlined in the preceding section, it is to be expected that in
case of a non-oriented reference object (tree) the deictic interpretation is chosen, in case of an
oriented reference object (yellow beetle) the intrinsic interpretation is preferred. The deictic
interpretation corresponds to the choice of subspace 1 for vor [in front oJ] and of subspace 3
for hinter [behind]; the intrinsic interpretation corresponds to the opposite assignments of
subspaces to prepositional phrases, i.e. subspace 3 for vor [in front oJ] and subspace 1 for
hinter [behind]. Ambiguity is only expected, if at all, in case of the intrinsically oriented yellow
beetle being the reference object. The results are shown in Table 1.
Table 1. Frequency of choices of either subspace when inter-
preting vor or hinter in an informal situation (lift home).
The left part of the table shows the frequencies for the
non-oriented reference object (tree), the right part for
the oriented object (yellow beetle).
driving test, receiving instructions by the examiner to park either in front of or behind the
reference object. Results are shown in Table 2.
Table 2. Frequency of choices of either subspace when inter-
preting vor or hinter in an institutionalized situation
(driving test). The left part of the table shows the
frequencies for the non-oriented reference object (tree),
the fight part for the oriented object (yellow beetle).
Here we find quite the opposite results, compared to the informal situation reported before:
Again, the orientedness of the reference object is an important determinant of the
interpretation. However, here we find predominating intrinsic interpretations when the
reference object is oriented, whereas the ambiguity arises with the non-oriented reference
object where the speaker can not guide the hearer's reaction by the use of one of both
prepositions, i.e., the prepositions do not differentiate between the alternative subspaces. Each
result has been replicated for several times and, thus, proved stable and reliable.
Summarizing the results obtained from German (Table 3), we find that an explanation of the
interpretation patterns of vor and hinter calls for the interaction of two factors: the formal or
informal characteristic of the social situation in which spatial communication occurs, and the
existing or non-existing intrinsic orientedness of the reference object.
Table 3. Two determinants of the interpretation of German vor
and hinter: the social situation and the orientedness of
the reference object.
reference object
situation
oriented non-oriented
P o s t hoc, the results may be explained as follows: In an institutionalized and formal situation,
people in our culture prefer a neutral reference point, an embedding of the spatial relation that
is not bound to the unstable position of the interlocutors, i.e., they prefer the intrinsic frame of
COMPREI IENSION Ol; SPATIAl, PREPOSITIONS 29
reference and they expect, in the role of the hearer, the intrinsic flame to be used. In an
informal situation, people prefer one of the involved individuals to be the reference point, i.e.,
they prefer the deictic system. The non-oriented tree in the lift-home situation and the oriented
car in the driving-test situation fit these standards--and the interpretation of spatial
prepositions is unambiguous. However, in the lift-home situation there is a conflict between the
intrinsic orientedness of the yellow beetle and the preference for the deictic flame; therefore,
deictic or intrinsic interpretations result alternatively. In the driving-test situation, there is a
conflict between the non-ofientedness of the tree and the preference for the intrinsic frame;
again, deictic or intrinsic interpretations result alternatively. In this case, the preference for the
non-personal embedding of the spatial relation becomes realizable by assigning the tree an
'objective' orientation on the basis of the extrinsic frame of reference. Moreover, the
interpretation patterns of vor and hinter are not symmetrical: vor is more often understood
deictically, hinter is more oi~en conceived intrinsic.
In further experiments not reported here (Grabowski, i. prep.; WeiB, 1993), we proved that the
observed inconsistencies of interpretation are not merely due to the fact that it would not
matter in the given situations where exactly to stop the car.
Consider two additional results: When, in the lift-home situation, the alignment of the oriented
reference object (the yellow beetle) with the direction of the road is removed (a difference of
just 30 ° is enough), the conflict disappears and unanimous (deictic) interpretations arise (see
Fig. 4).
I i~ca r
;'ub~ yellow
.... beetle
subspace • "I" " i ~1
lla sub,aceI~ car
*vor" 10 0
subspace
1 I 3
ii',ii =
:u'.
"vor" 8 2
In order to test the aforementioned typological hypothesis, it is necessary to study the use of
spatial prepositions in languages which have a prepositional system different from German.
With respect to the temporal axis and the primary horizontal axis, a language can provide their
users with a maximum of four different expressions as it is the case in French (devant/derri~re,
avant/apr~s), in Italian (davanti/dietro, prima/dopo) and in English (in front of~behind,
before~after). We will call these languages four preposition languages. German is a three
preposition language (vor/hinter, vor/nach), the same holds for Dutch (voor/achter, voor/na).
There are also two preposition languages like Latin where ante and post are used to express
spatial as well as temporal relations. We do not know of a one preposition language where
one single expression would differentiate neither between the temporal and the spatial system
nor between their marked and unmarked poles.
In order to show that there is in fact an influence of language topology and not merely a
phenomenon within one particular language, it is necessary to study more than one language of
each type. Therefore, we conducted the German experiments reported in the previous section
also in Dutch, French, Italian, and English. Thus, we contrasted three preposition languages
with four preposition languages (Fig. 6). The experimental instructions for the participants
have been translated into each target language; vor and hinter have been replaced by the
prepositions that are listed in the left half of the diagrams shown in Fig. 6. Instructions were
given by native speakers of the examined languages who served as experimenters; the
experiments were carded out at universities in the respective countries (for details see the
acknowledgments at the end of our paper).
French
devant avant
derri6re apr~s
German Italian
vor vor davanti prima
Dutch English
voor voor in front of before
==
achter na behind after
G /oriented :r°:nted
lift home intr.
driving | intr t'rising n,r
test / '
F orientednon-
lift home Intr. o~d
ambiguous driving
test intr ntr.
intended subspace was related to the reference object in the center spot by means of in front of
or behind, or vor and hinter, respectively. That is to say that speakers, when localizing, use
just those kinds of utterances that they can not unambiguously interpret when they take the
role of the hearer; they are not aware of the potential ambiguity and of the alternative frames of
spatial reference---which they master perfectly well. Even in the face of everyday
communicative fuzziness and ambiguity, they are convinced of the fact that language use is
clear and unambiguous (cf. Buhl, 1995). However, some of our participants contacted us after
months and reported that they lost their referential innocence, being always aware that
whenever something is in front of an object it might also be behind it at the same time. In this
respect, our experiments can, unintendedly, be considered a contribution to linguistic
enlightenment.
RESU
There are three conclusions that appear to be most important:
1. In Anglo-American research, the preference of the intrinsic system in cases where the
reference object is inherently oriented is otten claimed. This claim seems to be correct for the
type of language English belongs to. However, intrinsic preference is not a general principle.
For German, for example, the theoretical assumption is valid only within restrictions, because
of the fact that the influence of the prepositional inventory provided by a language on the
interpretation of dimensional prepositions is being overlooked.
2. The intrinsic orientedness of the reference object has already been considered in the
approaches that have been developed hitherto. However, there are at least two further
determinants of the use of dimensional prepositions: the characteristic of the social situation in
which spatial relations are conceived, and the prepositional inventory provided by the language
at issue.
3. In all languages we examined empirically, ambiguity occurs in systematic cases. It is not the
case that one language is generally more suitable than another language to express spatial
reference by the use of dimensional prepositions. However, according to the prepositional
inventory communication succeeds differently well in different cases.
ACKNOWLEDGMENTS
Our experiments in languages other than German were made possible by the hospitality and the
generous support from several colleagues: Wolfgang Klein (Nijmegen), Michel Fayol and
Bratrice Bourdin (Dijon), Bianca de Bernardi (Verona), George Miller and Christiane
Fellbaum (Princeton). We are grateful to Irene ter Haar, Alexa Heintze, Emanuela Antolini and
Shaft Landes who patiently and reliably instructed our participants in their native languages.
The English experiments were supported by a grant of the German Science Foundation (DFG)
to the first author. Gisela Zifonun contributed helpful comments on an earlier version of our
paper.
REFERENCES
Abkarian, G. G. (1982). Comprehension of deictic locatives: the object "behind" it. Journal of
Psycholinguistic Research, 11,229-245.
Buhl, H. (1995). Zur Blickpunktbezogenheit sprachlicher Lokalisationen: Der Genese-Effekt.
Unpublished Doctoral Thesis, University of Mannheim.
Carlson-Radvansky, L. A. and D. E. Irwin (1993). Frames of reference in vision and language:
Where is above? Cognition, 46, 223-244.
34 J. GRAI3OWSKIand 17.WEISS
Clark, H. H. (1973). Space, time, semantics, and the child. In: Cognitive Development and the
Acquisition of Language (T. E. Moore, ed.), pp. 27-63. Academic Press, New York.
Ehdch, V. (1985). Zur Linguistik und Psycholinguistik der sekund~ren Raumdeixis. In: Spra-
che und Raum. Psychologische und linguistische Aspelae der Aneignung und Verarbeitung
yon R~iumlichkeit (H. Schweizer, ed.), pp. 130-161. Metzler, Stuttgart.
Ehfich, V. (1989). Die temporale Festlegung lokaler Referenz. In: Raumkonzepte in Verste-
hemprozessen. Interdisziplina're Beitra'ge zu Sprache und Raum (Ch. Habel, M Herweg
and K. Rehk~mper, eds.), pp. 1-16. Niemeyer, Tfibingen.
Friederici, A. D. (1989). Raumreferenz unter extremen perzeptuellen Bedingungen: Perzep-
tion, Reprfisentation und sprachliche Abbildung. In: Raumkonzepte in Verstehensprozessen.
Interdisziplina're Beitrage zu Sprache und Raum (Ch. Habel, M. Herweg and K. Rehkam-
per, eds.), pp. 17-36. Niemeyer, Ttibingen.
Gosztonyi, A. (1976). Der Raum. Geschichte seiner Probleme in Philosophie und Wissen-
schaften. Alber, Freiburg.
Grabowski, J. (1994). Kommunikative Unsch~irfen - Zur Rezeption und Produktion von Rich-
tungspr~ipositionen am Beispiei von "vor" und "hinter". In: Sprache und Kognition:
Perspektiven moderner Sprachpsychologie (H.-J. Kornadt, J. Grabowski and R. Mangold-
Allwinn, eds.), pp. 183-208. Spektrum Akademischer Verlag, Heidelberg.
Grabowski, J. (i. prep.). Die Psychologie von "vor" und "hinter" - Produktion, Rezeption,
Kommunikation. Habilitation Thesis, University of Mannheim
Grabowski, J., Th. Herrmann and P. WeiB (1993). Wenn "vor" gleich "hinter" ist - zur multi-
plen Determination des Verstehens von Richtungspr~tpositionen. Kognitionswissenschafl, 3,
171-183.
Grabowski, J. and G. A. Miller (i. prep.). Factors affecting production and comprehension of
spatial prepositions: A comparison of German and American English.
Habel, Ch. (1989). "zwischen"-Bericht. In: Raumkonzepte in Verstehensprozessen. lnter-
disziplin~ire Beitra'ge zu Sprache und Raum (Ch. Habel, M. Herweg and K. Rehkamper,
eds.), pp. 37-69. Niemeyer, TObingen.
Herrmann, Th. (1990). Vor, hinter, rechts und links: das 6H-Modell. Psychologische Studien
zum sprachlichen Lokalisieren. Zeitschrift fiir Literaturwissenschaft und Linguistik, 78.
117-140.
Herrmann, Th. and J. Grabowski (1994). Sprechen -Psychologie der Sprachproduktion.
Spektrum Akademischer Verlag, Heidelberg.
Herskovits, A. (1986). Language and Spatial Cognition. An Interdisciplinary Study of the
Prepositions in English. Cambridge University Press, Cambridge.
Hill, C. A. (1982). Up/down, front/back, left/right. A contrastive study of Hausa and English.
In: Here and There. Cross-linguistic Studies on Deixis and Demonstration (J. Weissenborn
and W. Klein, eds.), pp. 13-42. Benjamins, Amsterdam.
Jackendoff, R. S. (1983). Semantics and Cognition. MIT Press, Cambridge.
Jackendoff, R. S. and B. Landau (1991). Spatial language and spatial cognition. In: Bridges
Between Psychology and Linguistics: A Swarthmore Festschrift for Lila Gleitman (D. J.
Napoli and J. A. Kegl, eds.), pp. 145-169. Erlbaum, Hillsdale.
Klein, W. (1994). Keine Kanguruhs zur Linken - Uber die Variabilitat von Raumvorstellungen
und ihren Ausdruck in der Sprache. In: Sprache und Kognition: Perspektiven moderner
Sprachpsychologie (H.-J. Kornadt, J. Grabowski and R. Mangold-Allwinn, eds.), pp. 163-
182. Spektrum Akademischer Verlag, Heidelberg.
Landau, B. and R. S. Jackendoff (1993). "What" and "where" in spatial language and spatial
cognition. Behavioral and Brain Sciences, 16, 217-265.
Levelt, W. J. M. (1982). Cognitive styles in the use of spatial direction terms. In: Speech,
Place, and Action. Studies in Deixis and Related Topics (R. J. Jarvella and W. Klein, eds.),
pp. 251-266. Wiley, Chichester.
LeveR, W. J. M. (1984). Some perceptual limitations on talking about space. In: Limits in Per-
ception. Essays in Honour ofMaarten A. Bouman (A. J. van Doom, W. A. van de Grind
and J. J. Koenderink, eds.), pp. 323-358. VNU Science Press, Utrecht.
Levelt, W. J. M. (1986). Zur sprachlichen Abbildung des Raumes: Deiktische und intrinsische
Perspektive. In: Perspektiven auf Sprache. Interdisziplinare BeitrgT"ge zum Gedenken an
HansHOrmann (H. G. Bosshardt, ed.), pp. 187-211. de Gruyter, Berlin.
COMPREHENSION OF SPATIALPREPOSITIONS 35
Linde, C. and W. Labov (1975). Spatial networks as a site for the study of language and
thought. Language, 51,924-939.
Miller, G. A. and Ph. N. Johnson-Laird (1976). Language and Perception. Cambridge
University Press, Cambridge.
Retz-Schmidt, G. (1988). Various views on spatial prepositions. AIMagazine, 9, 95-105.
Ruhrberg, P. and H. Rutz (1990). Raumliches Wissen und Semantik im Kontext der
Generierung von Wegbeschreibungen. In: Reprasentation und Verarbeitung raumlichen
Wissens (C. Freksa and Ch. Habel, eds.), pp. 235-249. Springer, Berlin.
Siegel. S. (1988). Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill, New
York.
Stutterheim, Ch. von and M. Carroll (1993). Raumkonzepte in Produktionsprozessen.
Kognitionswissenschaft, 3, 70-82.
Traugott, E. C. (1978). On the expression of spatio-temporal relations in language. In:
Universals of Human Language (J. H. Greenberg, ed.), Vol. 3: Word Structure, pp. 369-
400. Stanford University Press, Stanford.
Vater, H. (1991). Einf~hrung in die Raum-Linguistik (KLAGE K61ner Linguistische
Arbeiten--Germanistik; 24) (2rid ed.). Gabel, Hiarth-Efferen.
Weil3, P. (1993). Psychologische und sprachtvpologische Determinanten der Rezeption und
Interpretation von "vor" und "hinter". Unpublished Diploma Thesis, University of
Mannheim.
Wunderlich, D. (1981). Linguistic strategies. In: A Festschriftfor Native Speaker (F. Coulmas,
ed.), pp. 279-296. Mouton, The Hague.
Wunderlich, D. (1982). Sprache und Raum. Studium Linguistik, Heft 12/13_, 1-19/37-59.
Wunderlich, D. (1985). Raum, Zeit und das Lexikon. In: Sprache und Raum. Psychologische
und linguistische Aspekte der Aneignung und Verarbeitung von Raumlichkeit (H.
Schweizer, ed.), pp. 66-89. Metzler, Stuttgart.
Wunderlich, D. and M. Herweg (1991). Lokale und Direktionale. In: Semantics. An
International Handbook of Contemporary Research (A. von Stechow and D. Wunderlich,
eds.), pp. 758-785. de Gruyter, Berlin.